Biological functions of the autophagy-related proteins Atg4 and Atg8 in Cryptococcus neoformans

Autophagy is a mechanism responsible for intracellular degradation and recycling of macromolecules and organelles, essential for cell survival in adverse conditions. More than 40 autophagy-related (ATG) genes have been identified and characterized in fungi, among them ATG4 and ATG8. ATG4 encodes a cysteine protease (Atg4) that plays an important role in autophagy by initially processing Atg8 at its C-terminus region. Atg8 is a ubiquitin-like protein essential for the synthesis of the double-layer membrane that constitutes the autophagosome vesicle, responsible for delivering the cargo from the cytoplasm to the vacuole lumen. The contributions of Atg-related proteins in the pathogenic yeast in the genus Cryptococcus remain to be explored, to elucidate the molecular basis of the autophagy pathway. In this context, we aimed to investigate the role of autophagy-related proteins 4 and 8 (Atg4 and Atg8) during autophagy induction and their contribution with non-autophagic events in C. neoformans. We found that Atg4 and Atg8 are conserved proteins and that they interact physically with each other. ATG gene deletions resulted in cells sensitive to nitrogen starvation. ATG4 gene disruption affects Atg8 degradation and its translocation to the vacuole lumen, after autophagy induction. Both atg4 and atg8 mutants are more resistant to oxidative stress, have an impaired growth in the presence of the cell wall-perturbing agent Congo Red, and are sensitive to the proteasome inhibitor bortezomib (BTZ). By that, we conclude that in C. neoformans the autophagy-related proteins Atg4 and Atg8 play an important role in the autophagy pathway; which are required for autophagy regulation, maintenance of amino acid levels and cell adaptation to stressful conditions.


Introduction
Cryptococcosis is a systemic mycosis caused by encapsulated basidiomycete yeasts of the genus Cryptococcus [1]. The highest incidence of the disease is caused by C. neoformans, a global opportunistic pathogen prevalent in immunodeficient patients, mainly HIV-positive are required for vegetative development, conidiation, and virulence in plant tissues [24,25]. In the rice blast fungus Magnaporthe oryzae, Atg4 interacts and processes Atg8, leading to accumulation of autophagic bodies in the vacuole by nitrogen starvation. In addition, Atg4 is fundamental for hyphal development, conidiation, appressorium formation, and pathogenesis [26]. In Ustilago maydis, a plant pathogenic fungus, Atg8 impacts on the accumulation of autophagic bodies in the vacuole, cell survival under carbon starvation condition, and teliospores formation [27]. Atg8 is also important for virulence and sporulation in the chestnut blight fungus, Cryphonectria parasitica, and may be a target for a mycovirus that causes a reduction in pathogenicity [28]. For the human pathogenic fungus C. neoformans, a variety of studies were carried out to characterize the genes associated with the autophagy pathway. Hu et al. [29] described the importance of gene VPS34 in the autophagy process and, using RNAi, knocked down the Atg8 expression and demonstrated that this protein is important for virulence in a mouse model of cryptococcosis. Oliveira et al. [30] showed that atg7 mutants have reduced autophagic body formation and decreased cell viability under nitrogen deprivation. Beyond defects in the autophagy pathway, the ATG7 gene in C. neoformans contributes to cellular size and survival in mice lungs. Gontijo et al. [31] reported that the predicted cargo protein aspartyl aminopeptidase-like (Ape4) is important for survival at 37˚C, cell size, antifungal susceptibility, capsule formation, survival within macrophages, and virulence in mice infected with C. neoformans. Also, these authors pointed out that of 33 genes involved in S. cerevisiae autophagy, C. neoformans bares in its genome only 21 of them, indicating that likely the autophagy in the latter may function slightly different from what has been described for the former. Furthermore, Ding et al. [32] found that C. neoformans ATG genes ATG1, ATG7, ATG8 and ATG9 are required to maintain amino acid levels in a nitrogen depleted condition, contribute to sensitivity to the proteasome inhibitor BTZ and affect the virulence in a murine model of infection. Hence, it has been established for Cryptococcus the importance of Atg-related proteins in different biological aspects, yet the molecular mechanisms involving the autophagy regulation have not been fully explored. Thus, the present study aimed to investigate the C. neoformans Atg4 and Atg8 proteins by analyzing their complementarity of function with homologous proteins of S. cerevisiae, role in the regulation of autophagy pathway, and impact of mutant strains in response to multi-stress conditions.

Fungal strains and culture conditions
The yeast strains used in this study are presented in S1 Table. The mutant strains were generated in the C. neoformans var. grubii (KN99α) and S. cerevisiae (BY4741) backgrounds. Cultures were commonly maintained on rich medium YPD (1% yeast extract, 2% peptone and 2% glucose). Synthetic dextrose (SD) medium was prepared either with or without ammonium sulfate and amino acids (Sigma). In order to investigate the direct protein-protein interactions by the yeast two-hybrid system, DDO (SD/+N, -Leu -Trp), QDO (SD/+N, -Leu -Trp -Ade -His), QDO/X (QDO supplemented with 40 μg/mL X-α-Gal) and QDO/X/A (QDO/X with 200 ng/mL Aureobasidin A) media were used, as indicated by the manufacturer (Clontech). Cultures were incubated at 30˚C or 37˚C, in a rotary shaker with 150 rpm when required.

In silico analysis of autophagy-related proteins Atg4 and Atg8
The C. neoformans ATG4 (CNAG_02662) and ATG8 (CNAG_00816) genes were identified by BLASTp similarity search against the C. neoformans H99 database (taxid: 235443) available at the NCBI (National Center for Biotechnology Information) and using the amino acid sequences of S. cerevisiae Atg4 and Atg8 as a query (SGD accession numbers YNL223W and YBL078C, respectively). The protein conserved domains and signatures were obtained using the CDD (Conserved Domain Database) platform available online at NCBI. The illustration of the protein domains was made using the IBS 1.0.3 software (CUCKOO Workgroup). Multiple alignment of homologous amino acid sequences was performed using the ClustalW program [33]. Phylogenetic analysis was conducted using the Neighbor-Joining method with 1,000 bootstrap replicates [34], using the software MEGA v. 10.0.5 [35].

Protein-protein interactions by yeast two-hybrid assay
The direct protein-protein interactions were performed using the Matchmaker 1 Gold Yeast Two-Hybrid System (Clontech, Takara Bio), following the manufacturer's recommendations. For all the experiments the strain Y2HGold and the vectors pGBKT7 (bait) and pGADT7 (prey), supplied with the kit, were used. The coding sequence of each gene of interest was amplified by PCR using the specific primers listed in S2 Table. C. neoformans ATG8 (CNAG_00816) cDNA was cloned into GAL4 DNA-binding domain vector pGBKT7, while ATG3 (CNAG_06892); ATG4 (CNAG_02662) and ATG7 (CNAG_04538) cDNAs were cloned into the GAL4 transcriptional activation domain vector pGADT7. Then, the cloned vectors were individually transformed into the strain Y2HGold, according to the Yeastmaker™ Yeast Transformation System 2 (Clontech, Takara Bio) protocol, selected on SD/-Leu or SD/-Trp, and tested for the autoactivation. The activation of the reporter genes was evaluated by plating the cells on selective medium supplemented with 40 μg/mL X-α-Gal and 200 ng/mL Aureobasidin A. After autoactivation analysis, each pair of bait (pGADT7) and prey (pGBKT7) was cotransformed in Y2HGold and the transformants were selected on DDO medium. In addition, two independent clones were used to verify the interaction between bait and prey. Colonies growing in DDO were streaked onto QDO, QDO/X and QDO/X/A plates. The growth was observed after 5 days of incubation at 30˚C. The positive interactions were detected by cell survival in selective media that activate the reporter genes (AUR1-C, ADE2, HIS3, and MEL1) after bait and prey interaction. For this experiment, the plasmids pGBKT7-53 and pGADT7-T were used as a positive control, and pGBKT7-Lam and pGADT7-T were used as a negative control.

Yeast complementation and starvation challenge
For the yeast complementation assay we amplified by PCR the coding sequence of ATG4 (CNAG_02662) and ATG8 (CNAG_00816) C. neoformans genes, using the primers indicated in S2 Table. The coding sequence fragments were cloned into the HindIII and XbaI restriction sites of the yeast expression vector pYES2 (Invitrogen) using the Gibson Assembly 1 Master Mix (New England BioLabs). The recombinant plasmids were introduced into the S. cerevisiae (BY4741) atg4Δ and atg8Δ null mutants, whose construction is described below, following the lithium acetate (LiAc)/single-stranded DNA (SS-DNA)/polyethylene glycol (PEG) transformation protocol [36]. The transformants were selected by uracil prototrophy. The functional complementation effect was tested under nitrogen starvation as proposed by Matsuura et al. [37] with some modifications. First, the cells were cultured in YPD broth for 2 days at 30˚C with 150 rpm. Then yeast cells were washed with sterile saline 0.9%, transferred to liquid SD/-N/-AA (2% glucose and 0.17% yeast nitrogen base with neither amino acids nor ammonium sulfate) and incubated for a period of 20 days at 30˚C with 150 rpm in a rotary shaker. Lastly, the culture suspension had the OD 530 adjusted to 0.6 and the suspension had the concentration reduced in a 10-fold serial dilution. Five microliters of each serial dilution were spotted on YPD plates and incubated for 48 hours at 30˚C, before comparing the growth ratio between the wild type, mutant and/or complemented strains. The starvation challenge for C. neoformans was also performed using the nitrogen starvation protocol, in order to check the starvation-sensitive phenotype in the autophagy-related mutant strains. All the experiments independently were repeated three times.

Gene deletion and fungal transformation
The S. cerevisiae atg4Δ and atg8Δ mutants were generated by substituting the ATG4 (YNL223W) and ATG8 (YBL078C) coding regions with the KanMX6 selectable marker, which confers resistance to G418 (Geneticin). The gene replacement was performed by a homologous recombination strategy, using the primers listed in S2 Table. The selective marker was overlap-PCR amplified from plasmid pFA6-S65TGFP-KanMX6. The PCR fragments were introduced into S. cerevisiae BY4741 strain using the LiAc/SS carrier DNA/PEG heat shock transformation methodology [36]. The transformants were selected on YPD plates containing 200 μg/mL G418 and the deletion was confirmed by diagnostic PCR. The C. neoformans gene deletions were constructed by replacing, via homologous recombination, the ATG4 (CNAG_02662) and ATG8 (CNAG_00816) coding regions by a selectable marker hygromycin phosphotransferase. Sequences of all the oligonucleotides primers designed for gene-disruption are indicated in S2 Table. Briefly, the selectable marker was amplified from vector pPZP-Hyg2 (hygromycin resistance cassette) (donated by Dr. J Heitman [38]) by overlap-PCR extension, based on doublejoint PCR methodology [39]. Whereas, the flanking regions of a target gene to be deleted were PCR amplified using the KN99α gDNA as template. The generated fragments were introduced in wild type strain (KN99α) by biolistic transformation using the Biolistic 1 PDS-1000/He Particle Delivery System (BioRad) [40]. Subsequently, the mutants were plated and selected on YPD medium supplemented with 200 μg/mL hygromycin. After the gene deletion, reconstituted strains were generated by PCR amplification of the wild type native locus, that was biolistically delivered back into the mutant, co-transformed with the plasmid pPZP-Neo1 (G418 resistance cassette) (donated by Dr. J Heitman [38]). Reconstituted transformants were selected on YPD with 200 μg/mL G418. The gene replacements were tested by diagnostic PCR and confirmed by Southern blot analysis. The yeast genomic DNA extraction was performed following a previously described methodology [41] and the Southern blot technique was performed according to the protocol proposed by Sambrook et al. [42].

GFP fusion cassette and fluorescence microscopy
The C. neoformans ATG8 coding region was amplified by PCR and inserted into the BamHI and SpeI cloning sites of the vector pCN50 [43], resulting in a plasmid with the gene fused in frame to the C-terminus of GFP (green fluorescent protein). The primers to generate the fragments are presented in S2 Table. The cloned vectors were introduced into wild type KN99α and atg4 mutant strains by biolistic transformation. The selection of transformants was carried out in YPD agar containing 200 μg/mL G418 (Geneticin). The cellular localization of autophagy-related protein Atg8 was evaluated by fluorescence microscopy. In general, the yeast cells expressing the GFP fused proteins were grown overnight in liquid YPD at 30˚C with 150 rpm. The cells were harvested by centrifugation and washed twice in a phosphate-buffered saline (PBS). Then, cells were inoculated to a final concentration of OD 600 1.0 in one milliliter of YPD and SD/-N/-AA. Each condition was incubated for 4 hours at 30˚C and 37˚C in a rotary shaker. For the vacuole membrane staining, 1 μg/mL of FM4-64 dye (Invitrogen) was added to the culture after 3 h of incubation [44]. The samples were analyzed using an Olympus BX51 fluorescent microscope with an Olympus DP73 digital camera (Olympus). A total of one hundred cells per treatment distributed in different visual fields were evaluated. All acquired images were captured with 100× objective lens and processed with the program cellSens Dimension 1.16 (Olympus). Microscopy analyses were done in biological triplicate. The data were statistically analyzed in GraphPad Prism 7.0 software using two-way ANOVA, to compare different groups with more than one variable.

Protein extraction and western blotting
For the total protein extract preparation, yeast cells were first cultured in YPD broth until they reached the mid-log phase at 30˚C. Autophagy was then induced by transferring the cells to liquid SD without ammonium sulfate and amino acids. Nitrogen starvation induction was performed from 0-4 hours at 30˚C and 150 rpm. After induction, the yeast cells were harvested by centrifugation and resuspended in 500 μL lysis buffer (50 mM Tris pH 7.4, 100 mM NaCl, 0.5 mM PMSF, 10 mM orthovanadate, 50 mM NaF, 0.5% NP-40) containing protease and phosphatase inhibitors (Sigma-Aldrich). Glass beads were added to the cellular suspension and cells were lysed using a Mini-Beadbeater-16 (BioSpec). The protein concentration was determined by Bradford assay [45]. For western blotting, 50 μg of total protein were separated by SDS-PAGE on 10% polyacrylamide gels and transferred to nitrocellulose membranes (Bio-Rad) using the Trans-Blot 1 SD Semi-Dry Transfer Cell (Bio-Rad) at 15 V for 1 hour. Then, the membranes were blocked in 5% skimmed milk in Tris-buffered saline (TBS) containing 0.1% Tween 20. Membranes were further incubated with primary mouse anti-GFP (Thermo Fisher Scientific) or rabbit anti-histone (Cell Signaling) in a 1:2,000 dilution. Following this, the membranes were incubated with anti-mouse IgG or anti-rabbit IgG (Cell Signaling) secondary antibodies conjugated to horse-radish peroxidase in a 1:5,000 dilution. The immunoreactive bands were detected using SuperSignal™ West Pico Chemiluminescent Substrate (Thermo Fisher Scientific) and visualized using ImageQuant LAS (GE Healthcare). This experiment was carried out three times independently.

Stress sensitivity assay
The procedure to evaluate yeast growth under stress conditions was performed by incubating overnight the C. neoformans strains in 5 mL YPD broth at 30˚C in a rotary shaker. The cells were precipitated by centrifugation, washed twice in sterile saline and diluted to a concentration of 2 × 10 6 CFU/mL and 10-fold diluted until to 2 × 10 2 CFU/mL. Five microliters of each dilution were spotted on plates containing the stress-inducing agents. For the stress analysis, YPD agar was supplemented with 0.5%-0.75% Congo Red (cell wall integrity), 0.03%-0.04% SDS (cell membrane integrity), pH adjusted from 6 to 8 (alkaline stress), 0.75-1 M KCl (osmotic stress), 0.75-1 M NaCl (saline stress), 2.5-4 mM H 2 O 2 (oxidative stress), and 1-2.5 mM NaNO 2 (nitrosative stress); furthermore, SD agar was supplemented with different combinations of ammonium sulfate and amino acids (starvation stress). The spotted cells were incubated at 30˚C and 37˚C for a period of 2 days and then photographed. Each condition was repeated in duplicate.

Sensitivity to the proteasome inhibitor Bortezomib (BTZ)
The growth inhibition effect of Bortezomib (BTZ) upon C. neoformans strains (wild type and mutants) was tested as previously described by Ding et al. [32]. Briefly, an isolated colony was inoculated in 5 mL YPD and incubated overnight at 30˚C with 150 rpm. Cells were harvested by centrifugation, washed three times in a phosphate-buffered saline (PBS) and diluted to a final concentration of 10 6 cells/mL in liquid SD/-N/-AA, supplemented or not with 50 μg/mL BTZ (Accord/Intas Pharmaceuticals). The culture was incubated for a period of 48 hours at 30˚C in a rotary shaker. To determine the fold change of the initial CFU, cells were plated at 0 and 48 h on YPD and incubated at 30˚C for 48 hours. The CFU was calculated and the data were statistically analyzed in GraphPad Prism 7.0 software using unpaired t test for comparison of two groups. The experiment was performed in two independent biological replicates.

Identification of ATG8 in C. neoformans
We initiated our study by identifying the ATG8 orthologue gene in the genome of C. neoformans. The amino acid sequence of S. cerevisiae Atg8 protein (SGD accession number: YBL078C) was used as a query for the BLASTp similarity search against the Cryptococcus neoformans H99 database (taxid: 235443) available at the NCBI. The result revealed a putative microtubule binding protein containing 126 amino acids with 78% identity to the S. cerevisiae Atg8 protein, encoded in C. neoformans by the CNAG_00816 gene on chromosome 1 and confirming the identification previously published by other authors [29,31]. Next, we performed the phylogenetic analysis of multiple alignment using eight protein sequences with homology to Atg8 belonging to basidiomycetes and ascomycetes species. We verified that the Atg8 protein of C. neoformans is grouped in the same cluster of C. gattii and, moreover, C. neoformans Atg8 is phylogenetically closer to other ascomycetes species such as Neurospora crassa, Aspergillus flavus and Magnaporthe oryzae than S. cerevisiae (S1 Fig). Nevertheless, when comparing the structural sequence of different Atg8 homologous proteins between those same fungal species, we verified the presence of some conserved domains (S2 Fig). The data using the CDD platform (Conserved Domain Database, NCBI) showed the presence of the Ubiquitin-like superfamily (UBQ) domain and predicted Atg4 and Atg7 proteins binding sites (filled and unfilled arrows, respectively). In addition, it was possible to verify the presence of the conserved glycine residue, necessary for the cleavage of the C-terminal portion by the cysteine protease Atg4, allowing the subsequent conjugation of Atg8 with PE [13] (S2 Fig). These results indicate that C. neoformans Atg8 is highly conserved with orthologues from different fungal species.

Interactions of Atg8 with autophagy-related proteins Atg3, Atg4 and Atg7
It is known that in S. cerevisiae Atg8 binds directly to other autophagy-related proteins such as Atg3, Atg4 and Atg7, which are responsible for the processing of Atg8 to a form required for autophagosome vesicle synthesis [46,47]. Since C. neoformans Atg8 has predicted binding sites for Atg4 and Atg7 (S2 Fig), we investigated whether in C. neoformans there would also be direct interaction of Atg8 with these proteins, including Atg3. Direct protein-protein interactions were performed by yeast two-hybrid analysis. C. neoformans full-length cDNA sequences were used to construct both bait (pGBKT7) and prey (pGADT7) plasmids with the Match-maker™ Gold Yeast Two-Hybrid System (Clontech). The results of the two-hybrid experiments performed with two independent clones demonstrated that, individually, bait and preys do not autoactivate the reporter genes (S3 Fig); however, when co-transformed, the bait pGBKT7-ATG8 interacted with all prey containing ATG3 (CNAG_06892), ATG4 (CNAG_02662) or ATG7 (CNAG_04538) sequences in all selective media (Fig 1). These observations suggest that Atg3, Atg4 and Atg7 are proteins that, by interacting with Atg8, may be involved in the autophagy pathway in C. neoformans.

Characterization of autophagy-related protein Atg4 in C. neoformans
Since the Atg8 protein in C. neoformans shows the glycine residue (Atg8 G117 ) in the C-terminus portion (S2 Fig) and interacts directly with Atg4 via two-hybrid system (Fig 1), we decided to include in our study the analysis of Atg4. In order to identify the ATG4 orthologous gene, we used the amino acid sequence of S. cerevisiae Atg4 (SGD: YNL223W) as a query. The search in the BLASTp algorithm against the genome database of C. neoformans H99 (taxid: 235443) revealed similarity with the gene CNAG_02662 located in chromosome 3 and with 33% identity with the Atg4 of S. cerevisiae. In C. neoformans, ATG4 encodes a putative cysteine protease with 1,185 amino acids. As observed in S. cerevisiae, Atg4 from C. neoformans has a conserved domain of the Peptidase_C54 Superfamily between the amino acids 480-982: this domain is present in a group of proteolytic proteins capable of hydrolyzing the peptide bond. Moreover, when aligning the multiple sequences of Atg4 homologous proteins, we observed the catalytic residue of cysteine, which is conserved in all fungi species analyzed (S4 Fig). These results suggest that the functions of Atg4 in Saccharomyces and Cryptococcus may be conserved.
ATG4 and ATG8 of C. neoformans functionally complement the atg4 and atg8 mutants of S. cerevisiae As we observed that Atg proteins Atg8 and Atg4 are highly conserved in different fungal species, we next performed an assay to verify the hypothesis that the autophagy-related proteins Atg4 and Atg8 of Cryptococcus have the same functions as their well-studied counterparts in Saccharomyces. For this study, heterologous complementation in S. cerevisiae mutants was tested. First, we obtained the atg4Δ and atg8Δ null mutants by homologous recombination in S. cerevisiae (S5 Fig). Then, the full-length cDNA sequences of C. neoformans Atg4 and Atg8 were cloned into the expression vector pYES2 (Invitrogen) and transformed, respectively, into the atg4Δ and atg8Δ mutants of S. cerevisiae. After 20 days of culture in SD/-N/-AA broth medium, it was possible to observe a nutritional deprivation sensitivity phenotype for the S. cerevisiae atg4Δ and atg8Δ mutants. However, when these mutants were complemented with the Cryptococcus proteins, we observed that CnATG4 and CnATG8 recovered their growth defects, restoring the cell survival on nitrogen-lacking medium to the wild type strain level (Fig 2). Therefore, these results indicate that Atg4 and Atg8 of C. neoformans are functionally homologous with the S. cerevisiae autophagy-related proteins.

Involvement of proteins Atg4 and Atg8 in the autophagy pathway
In S. cerevisiae the ATG gene deletions leads to mutants with sensitivity to nitrogen starvation [48]. Based on that, we decided to investigate the response to starvation for ATG4 and ATG8 disrupted genes in C. neoformans. For this study, we first obtained individuals atg4 and atg8 mutants using the double-joint methodology (S6A and S6C Fig). After confirming the gene deletion by Southern blot analyses (S6B and S6D Fig), we reconstituted the ATG8 gene into the atg8 mutant by PCR amplification of the CNAG_00816 gene, which was co-transformed with the plasmid pPZP-Neo1 (G418 resistance). For the atg4 mutants we tested two independent strains, since we did not manage to restitute the wild type copy back into the mutant. For the starvation sensitivity experiment, we induced the autophagy in the wild type (KN99α), reconstituted (atg8+ATG8) and atg mutant strains by culturing the cells for 20 days in liquid SD/-N/-AA. After this period of starvation challenge, we observed the loss of cell viability for atg4 and atg8 mutants, compared to wild type and reconstituted phenotypes (Fig 3).

Atg4 processes Atg8 during nitrogen starvation
In order to elucidate in C. neoformans the processing of Atg8 involving the cysteine protease Atg4 upon autophagy induction, as described in S. cerevisiae [19], we introduced into KN99α and atg4 mutant strains the GFP-Atg8 plasmid. We next examined by western blot the vacuolar proteolysis of Atg8 under nitrogen starvation from a period of 0-4 h. When the wild type strain expressing GFP-Atg8 was incubated in SD/-N/-AA broth it was possible to visualize over time the bands of the GFP-Atg8 fused protein as well as the free GFP bands (Fig 4). The KN99α data suggest a normal autophagic flux, with Atg8 being degraded by releasing the GFP fragment in the course of time. On the other hand, in the atg4 mutant only the full length GFP-Atg8 bands were detected (Fig 4), suggesting that the autophagosome synthesis was blocked and, consequently, Atg8 was not degraded.

Atg8 intravacuolar localization occurs in an Atg4-dependent manner
In S. cerevisiae, the ATG4 gene knockout generates mutant cells defective in delivering Atg8 to the vacuole during the autophagy pathway, resulting in GFP-Atg8 diffuse distribution in the cytoplasm and, consequently, indicating that Atg8 processing by Atg4 is necessary for its subsequent vacuolar association [49]. To further understand the impact of ATG4 deletion in the mechanism of autophagy in C. neoformans, we next evaluated the Atg8 cellular localization by fluorescence microscopy, using the strains containing the GFP fused proteins. For that, we monitored the GFP-Atg8 localization upon cell cultivation in rich medium (YPD) and starvation medium (SD/-N/-AA) for 4 hours at 30˚C and 37˚C, after labeling the vacuole membrane with FM4-64 (Fig 5). In KN99α, the fluorescence visualization of cells growing in rich medium

PLOS ONE
Atg4 and Atg8 in Cryptococcus neoformans revealed that most of the GFP-Atg8 was distributed in the cell cytoplasm. However, after autophagy induction by starvation, we observed a significant translocation of GFP-Atg8 to the vacuolar lumen at 30˚C and 37˚C (Fig 5A). In contrast, for the atg4 mutants cultured in YPD, Atg8 fused with GFP was dispersed in the cytoplasm. On the other hand, when these cells lacking the protein Atg4 were cultured in SD/-N/-AA, we observed that GFP-Atg8 remains localized in the cytoplasm, demonstrating that Atg4 is required for the shifting of Atg8 to vacuole lumen (Fig 5A). When comparing the GFP localization according to the temperature, we observed a similar percentage of GFP-Atg8 in cells growing at 30˚C and 37˚C; in which for the wild type 86.3% and 82.3% of GFP-Atg8 were located in the cytoplasm (30˚C and 37˚C, respectively) and for atg4 mutant 93.3% and 93.7% were located intravacuolar (30˚C and 37˚C, respectively). The data were statistically significant only for the wild type strains, for which we observed for cells cultured in YPD an average of 93% of GFP localized in the cytoplasm and, for cells growing in SD/-N/-AA, we found around 84% of GFP-Atg8 localized intravacuolar (Fig 5B and 5C).
Together, these results support the idea that Atg4 and Atg8 proteins in C. neoformans are involved in an autophagic mechanism, since the lack of both proteins impairs yeast growth in an autophagy induced environment. Moreover, the ATG4 gene deletion affects the vacuolar breakdown of Atg8, indicating that Atg8 was not processed and could not form the autophagosome vesicle during the autophagy pathway. Additionally, the correct Atg8 intravacuolar localization occurs in an Atg4-dependent manner.

Requirement of autophagy-related genes during stress growth condition
Recently it was shown that in C. neoformans functional autophagy pathway is important to several biological events related to pathogenesis [30][31][32]. To better characterize the effect of a defective autophagy process, we performed experiments to evaluate the growth phenotype in multi-stress conditions for the wild type (KN99α), atg4 (two independent strains), atg8 and reconstituted atg8+ATG8. The stress conditions tested were those performed by many authors to access the effect of several mutations throughout the C. neoformans genome that are thought to be important for pathogenesis [50,51]. For this analysis, we could not observe any phenotypic difference for the mutants growing under starvation (SD+/-N+/-AA), alkaline (pH), osmotic (KCl), saline (NaCl), nitrosative (NaNO 2 ), and cell membrane integrity (SDS) stress (data shown in S7 Fig). However, compared to the wild type, it was possible to notice growth differences between atg4 and atg8 strains in response to YPD supplemented with H 2 O 2 and Congo Red, as shown in Fig 6. The ATG4 and ATG8 gene deletion resulted in an increased tolerance to hydrogen peroxide, supposedly affecting the fungal antioxidant defense and improving the yeast growth in the presence of H 2 O 2 . For the oxidative stress experiment, the atg4 and atg8 showed a growth improvement on YPD with 4 mM H 2 O 2 at 30˚C. When incubating the cells at 37˚C, we observed for the autophagy mutants an increment of resistance to hydrogen peroxide on YPD containing 3 mM H 2 O 2 and a slight cell growth on 4 mM H 2 O 2 , when compared to wild type strain ( Fig 6A). Furthermore, the atg4 and atg8 growth was impaired by incubating the cells on YPD supplemented with 0.65% and 0.75% Congo Red at 30˚C and 37˚C. Therefore, the disruption of the autophagy pathway somehow impairs C. neoformans to efficiently respond to this stressful condition (Fig 6B).
It is known that the autophagy and ubiquitin-proteasome pathways (UPP) help to maintain the amino acid homeostasis in the cell [52][53][54]. Ding et al. [32] described that disrupting the UPP by the proteasome inhibitor bortezomib (BTZ at 50 μg/mL) had an amplified effect upon autophagy mutants when growing under nitrogen starvation. Therefore, we decided to evaluate the impact of BTZ presence when growing atg4 and atg8 mutants under nitrogen depletion medium. As described previously by Ding et al. [32] and observed in our analysis, BTZ reduced the growth pattern of the wild type strain (KN99α) after 48 hours under nitrogen deprivation, but it was not a significant decrease in the CFU (Fig 6C). However, the growth impairment was rather pronounced for the atg8 mutant treated with BTZ, which had the growth reduced to 88% compared with the same strain without treatment. As expected, this impact was remediated when a wild type copy of ATG8 was reintroduced in the atg8 mutant, confirming the results published by Ding et al. [32]. Fig 6C shows that the two atg4 independent mutants have significantly reduced their proliferation by 88% and 93% in the presence of BTZ, compared with the same strain without BTZ. This experiment allows us to suggest that the lack of Atg4 and Atg8 impairs the progression of the autophagy mechanism; additionally, the deficiency of autophagy in combination with the inhibition of proteasome formation causes an unbalance in the amino acid homeostasis, leading to a drastic drop on nutrients recycling, that might impairs the cell to grow and/or survive under nitrogen deprivation.
Overall, these results indicate that in C. neoformans the autophagy mechanism may prepare the cell to respond to adverse environmental conditions by recycling nutrients likely, contributing to yeast survival.

Discussion
Due to the fact that C. neoformans causes infections mainly in immunodeficient individuals [55], affecting up to 223,000 individuals per year [56] and that there is a constraint regarding treatment; it is necessary to broad the knowledge on this fungus biology in order to search for alternatives for treatment and disease control. Understanding the molecular mechanisms that connect both pathogenicity and autophagy could provide us new insights into the relationship established during the fungal infection [57]. One of the central steps of the autophagy pathway is the autophagosome formation in which, in yeast model S. cerevisiae, Atg4 is important to start the autophagosome synthesis by processing Atg8 [58]. Therefore, we characterized in the present study the autophagy-related proteins Atg4 and Atg8 in C. neoformans, hoping to clarify new information about their roles in the biology of this pathogenic fungus.
The multiple protein sequence alignment revealed that Atg8 in C. neoformans contains the C-terminal glycine (Gly117) that is highly conserved in the fungi species analyzed in our study, and also reported in other organisms like Bipolaris maydis, Schizosaccharomyces pombe, Arabidopsis thaliana, Rattus norvegicus, Homo sapiens, etc [59,60]. Moreover, we found by two-hybrid analysis that C. neoformans Atg8 interacts with Atg3, Atg4 and Atg7. In S. cerevisiae, Atg8 is processed in the C-terminal region by Atg4, to expose the glycine residue; then processed Atg8 is activated by an E1-like enzyme Atg7 and, subsequently, transferred to the E2-like enzyme Atg3, to promote the Atg8 conjugation with PE (Atg8-PE). Atg8-PE is essential for the double membrane autophagosome formation [58]. Among the three confirmed interactions, we decided to better investigate the autophagy-related protein 4.
Atg4 is a cysteine protease of Clan CA and family C54 that is a crucial for both Atg8-PE conjugation and subsequent delipidation [61]. Atg4 was first identified in yeast by Lang et al. [46], who demonstrated that this protein was essential for the process of delivering the autophagic vesicles to the vacuole lumen in S. cerevisiae. Since then, Atg4 has been an important target to study autophagy in different eukaryotic organisms. In our analysis of Atg4 we observed that, as in S. cerevisiae, the C. neoformans cysteine protease has the conserved domain of Pepti-dase_C54 Superfamily. The protease database MEROPS [62] describes the family C54 as a group of endopeptidases that hydrolases a peptide bond, with specificity to glycyl bonds. Furthermore, in S. cerevisiae, the catalytic residue of cysteine (Cys147) was identified as responsible for initially processing Atg8 at the C-terminus, exposing the glycine residue and allowing Atg8 to conjugate with PE (Atg8-PE) [63]. Comparing the Atg4 multiple sequence alignment with fungal species, the C. neoformans putative catalytic residue of cysteine (Cys568) is conserved. Other authors studying alignments of Atg4 homologues found that the Cys residue was highly conserved among different species, beyond those presented in this study, including Botrytis cinerea, Candida glabrata, Fusarium graminearum, Penicillium oxalicum, Arabidopsis thaliana, Mus musculus, Homo sapiens, and others [20,24,26].
Evidence that Atg4 and Atg8 in C. neoformans are highly conserved led us to validate their functions by heterologous complementation in Saccharomyces. In fungi, functional complementation of autophagy genes in Saccharomyces mutants was demonstrated for Botrytis cinerea [24,25], Magnaporthe oryzae [26], and Sordaria macrospora [23]. Here, we analyzed the growth viability of the mutants and the complemented strains after cultivating for 20 days in starvation medium (SD/-N/-AA). In general, S. cerevisiae atg null mutants tend to lose their ability to survive in nitrogen starvation conditions [48]. In our study we observed that the S. cerevisiae atg4Δ and atg8Δ strains had defective growth in SD/-N/-AA. However, the phenotypes were recovered to the wild type level, once each mutant was complemented with their respective C. neoformans homologous gene. The same was observed when evaluating specifically the C. neoformans autophagy mutated strains, in that atg4 and atg8 also show a sensitivity phenotype after cultivated in depleted nutrient condition. In addition, Gontijo et al. [31] studying 21 putative genes involved in autophagy, showed that in C. neoformans the ATG4 and ATG8 gene transcription is induced by nitrogen starvation at 30˚C and 37˚C. Onodera and Ohsumi [52] reported that the loss of cellular viability in these atg mutants, cultured in nutritional deprivation, may be due to failure to sustain the amino acids supply, leading to inability to synthesize new proteins. Besides that, Suzuki et al. [64] presented that the cell death of atg mutants is in consequence of deficiency in respiratory function. The authors showed that, under starvation condition, the autophagy-defective mutants tend to accumulate reactive oxygen species (ROS), causing inhibition of mitochondrial DNA replication and resulting in cells unable to maintain mitochondrial function.
In most eukaryotic organisms, Atg8 is a central protein used to infer the progress of autophagy. Based on that, a reliable method to better understand the autophagy molecular mechanism is through monitoring the vacuolar delivery and proteolysis of GFP-Atg8 [65]. Here, we demonstrated that the fusion protein GFP-Atg8 can be also used to monitor the autophagy pathway in C. neoformans. Our western blot analysis revealed that Atg4 is essential for Atg8 vacuolar proteolysis, as the observations evidenced in yeast model S. cerevisiae [66]. Confirming our findings, the results obtained previously by other authors show that although Atg8 is expressed in the atg4 mutant cells, the protein cannot be processed in the C-terminal region to form the Atg8-PE conjugate, interrupting the autophagosomal membrane synthesis and, by that, blocking the GFP-Atg8 degradation in the vacuole lumen [66,67]. Furthermore, fluorescence microscopy analysis allowed us to observe the GFP-Atg8 flux during nitrogen starvation. We demonstrated in C. neoformans that GFP-Atg8 is localized inside the vacuole in an Atg4-dependent manner. As previously published for S. cerevisiae, in wild type cells growing in a normal autophagy condition, GFP-Atg8 can be visualized as a puncta close to the vacuole, indicating that Atg8 is located in the pre-autophagosomal structure (PAS), in which the Atgrelated proteins are recruited for autophagosome formation [68]. Although we could not see in C. neoformans a perivacuolar puncta signal for GFP-Atg8 when KN99α cells were cultured in starvation medium (SD/-N/-AA), we believe that the starvation induction for a period of 4 hours provided us with information on the later steps of the autophagy pathway. In this case, the GFP signal was concentrated within the vacuole, suggesting that the autophagosome double membrane was formed, fused with the vacuole, and lastly, the inner autophagosome membrane covered with GFP-Atg8-PE was delivered to the vacuole and degraded by vacuolar proteases [65].
Recently, new studies have reported that Atg proteins are involved in other biological process beyond autophagy. Oliveira et al. [30] showed that ATG7 deletion in C. neoformans impairs not only the autophagy mechanism, but also affects the cellular size, reduces the pathogenicity in infected larvae, impairs the colonization in mice lung, and decreases the cell resistance to fungicidal activity of neutrophils. Ding et al. [32] found that ATG1, ATG7, ATG8, and ATG9 genes in C. neoformans are involved in the autophagy mechanism and contribute to the virulence of the fungus. The authors observed that these mutants presented impaired maintenance of amino acids homeostasis under nitrogen starvation, reduced levels of virulence in a murine inhalation model, and contribute in different ways to the immune response and ability to promote the virulence. In our study, we found that both ATG4 and ATG8 deletion in C. neoformans resulted in cells that were more resistant to oxidative stress when cultivated with hydrogen peroxide, defective growth in the presence of Congo Red maybe as a result of a cell wall integrity damage, and sensitivity to the proteasome inhibitor BTZ.
Several studies have found evidence for an interaction between autophagy and ROS. Most of the data in the literature have shown that autophagy can be also controlled by ROS, which are responsible for the modulation of survival or cell death cascades, under oxidative inductions [69]. Experiments performed in Saccharomyces by Pérez-Pérez et al. [20] show that Atg4 is one of the autophagy-related proteins controlled by reactive oxygen species. The authors observed that Atg4 activity is inhibited by oxidation, in a H 2 O 2 concentration-dependent manner, thereby affecting the autophagic machinery. In most cases, basal levels of ROS play an important homeostatic role for cell survival but, at high levels, ROS are lethal to the cells [69]. For C. neoformans we observed that atg4 and atg8 mutants are tolerant to higher concentrations of hydrogen peroxide (3 and 4 mM H 2 O 2 ) compared to the wild type strain. It is important to note that a similar finding was obtained in mouse embryo fibroblasts (MEFs) cells: atg5 mutant cells were more resistant to cell death after exposure to H 2 O 2 , attributed by subsequent compensatory activation of extracellular signal-regulated kinases 1 and 2 (ERK1/2), which might interfere with cell survival against the oxidative damage [70].
Moreover, in C. neoformans, atg4 and atg8 mutants have an impaired growth in the presence of 0.65% and 0.75% Congo Red at 30˚C and 37˚C. Congo Red is cell wall-perturbing agent with high affinity for chitin, that causes weakness in the cell wall, activating the stress response [71]. For the endophytic fungus Harpophora oryzae, the ATG5 gene displays an important role in autophagy, sporulation and cell wall integrity. In the presence of Congo Red the growth is highly inhibited in the atg5 mutants, affecting the mycelial development [72]. Another finding in our investigation was the fact that ATG4 and ATG8 gene disruption resulted in sensitivity to the proteasome inhibitor BTZ. In accordance with our results, Ding et al. [32], studying the role of BTZ in the absence of a nitrogen source, reported that deletion of ATG1, ATG7, ATG8 and ATG9 genes in C. neoformans causes BTZ hypersensitivity cells, by inhibiting the ubiquitin-proteasome pathway. In addition, the authors pointed out that the autophagy mechanism is upregulated to compensate the proteasome inhibition; however, the effect of BTZ in ATG mutant cells is exacerbated when autophagy is compromised, resulting in a decrease in the availability of amino acids and other nutrients to the cell.
Taking these results together and based on Saccharomyces autophagy studies, we hypothesize the scenario that during the autophagy pathway in C. neoformans, Atg4 processes Atg8 by exposing the C-terminal glycine (Gly117). Thereby, processed Atg8 interacts with Atg7 and Atg3 to promote its conjugation with PE (Atg8-PE), allowing the double membrane autophagosome assembly/expansion, followed by the autophagosome-vacuole fusion and delivery of cargo to the vacuolar lumen. However, further experiments need to be developed to confirm our hypothesis. In conclusion, our study suggests that for C. neoformans the proteins Atg4 and Atg8 are involved in an autophagy mechanism under starvation induced conditions. The results also show that functional autophagy pathway may be part of distinct biological process in C. neoformans. Idnurm.